System and method for specifying batch execution ordering of requests in a storage system cluster
A system and method specifies batch execution ordering of requests in a cluster of storage systems or nodes. Each node is generally organized as a network element and a disk element. One or more sessions are established between the elements in the cluster. Each session comprises a plurality of channels disposed over connections, wherein each channel enables multiple requests to be sent over a connection. Each request is identified by a unique identifier (“request id”) that is generally defined as the combination of a channel number and a sequence number. The request id is extended to include a batch number that provides an execution ordering directive within a channel. All requests with the same batch number in a channel can be executed in arbitrary order or concurrently. Ordering is imposed only when the batch number changes, e.g., increases. Although more than one request in a channel can have the same batch number, all requests with the same batch number are executed before any request with a higher batch number.
1. A method for operating a computer data storage system, comprising:
connecting a plurality of network interfaces with a computer network to exchange messages with one or more clients connected to the computer network;
connecting a plurality of interfaces to data storage devices to a plurality of data storage devices;
connecting both the plurality of network interfaces and the plurality of interfaces to data storage devices to a cluster switching fabric;
establishing, through the cluster switching fabric, a communications session between a selected network interface of the plurality of network interfaces and a selected interface to a data storage device of the plurality of interfaces to data storage devices;
establishing for the session a plurality of connections through the cluster switching fabric between the selected network interface and the selected interface to a data storage device, each connection having established a conventional transport protocol;
receiving a plurality of requests from a client of the one or more clients, each request of the plurality of requests having assigned a unique sequence number, each request being an input/output request to a selected data storage device connected to the selected interface to a data storage device;
dividing the plurality of requests into a plurality of subsets of requests;
assigning a batch number to each subset of requests, the batch numbers increasing monotonically so that each subset of requests is assigned a unique batch number;
executing a first subset of requests in arbitrary order over the plurality of connections in response to a first batch number; and
executing a second subset of requests in response to a second batch number after execution of all of the first subset of requests have completed.
2. The method as in claim 1 , further comprising:
using a TCP/IP protocol for the conventional transport protocol.
3. The method as in claim 1 , further comprising:
using a writable storage device media as a storage device of the plurality of storage devices.
4. The method as in claim 1 , further comprising:
using an optical writable storage device as a storage device of the plurality of storage devices.
5. The method as in claim 1 , further comprising:
using a DVD writable storage device as a storage device of the plurality of storage devices.
6. The method as in claim 1 , further comprising:
using an electronic random access memory as a storage device of the plurality of storage devices.
7. The method as in claim 1 , further comprising:
using a micro-electro mechanical memory as a storage device of the plurality of storage devices.
8. The method as in claim 1 , further comprising:
using a magnetic disk as a storage device of the plurality of storage devices.
9. The method as in claim 1 , further comprising:
using a RAID array of magnetic disks as a storage device of the plurality of storage devices.
10. A computer data storage system, comprising:
a plurality of network interfaces connected with a computer network to exchange messages with one or more clients connected to the computer network;
a plurality of interfaces to data storage devices connected to a plurality of data storage devices;
a cluster switching fabric connected to both the plurality of network interfaces and the plurality of interfaces to data storage devices;
an operating system to establish, through the cluster switching fabric, a communications session between a selected network interface of the plurality of network interfaces and a selected interface to a data storage device of the plurality of interfaces to data storage devices;
the operating system to establish for the session a plurality of connections through the cluster switching fabric between the selected network interface and the selected interface to a data storage device, each connection having established a conventional transport protocol;
a selected network interface of the plurality of network interfaces to receive a plurality of requests from a client of the one or more clients, each request of the plurality of requests having assigned a unique sequence number, each request being an input/output request to a selected data storage device connected to the selected interface to a data storage device;
a first process to divide the plurality of requests into a plurality of subsets of requests;
a second process to assign a batch number to each subset of requests the batch numbers increasing monotonically so that each subset of requests is assigned a unique batch number;
the operating system to execute a first subset of requests in arbitrary order over the plurality of connections in response to a first batch number; and
the operating system to execute a second subset of requests in response to a second batch number after execution of all of the first subset of requests have completed.
11. The data storage system as in claim 10 , further comprising:
a TCP/IP protocol used for the conventional transport protocol.
12. The data storage system as in claim 10 , further comprising:
a writable storage device media is a storage device of the plurality of storage devices.
13. The data storage system as in claim 10 , further comprising:
an optical writable storage device is a storage device of the plurality of storage devices.
14. The data storage system as in claim 10 , further comprising:
a DVD writable storage device is a storage device of the plurality of storage devices.
15. The data storage system as in claim 10 , further comprising:
an electronic random access memory is a storage device of the plurality of storage devices.
16. The data storage system as in claim 10 , further comprising:
a micro-electro mechanical memory is a storage device of the plurality of storage devices.
17. The data storage system as in claim 10 , further comprising:
a magnetic disk is a storage device of the plurality of storage devices.
18. The data storage system as in claim 10 , further comprising:
a RAID array of magnetic disks is a storage device of the plurality of storage devices.
19. A computer readable storage media, comprising:
said computer readable storage media containing instructions for execution on a processor for a method of operating a computer data storage system, having,
connecting a plurality of network interfaces with a computer network to exchange messages with one or more clients connected to the computer network;
connecting a plurality of interfaces to data storage devices to a plurality of data storage devices;
connecting both the plurality of network interfaces and the plurality of interfaces to data storage devices to a cluster switching fabric;
establishing, through the cluster switching fabric, a communications session between a selected network interface of the plurality of network interfaces and a selected interface to a data storage device of the plurality of interfaces to data storage devices;
establishing for the session a plurality of connections through the cluster switching fabric between the selected network interface and the selected interface to a data storage device, each connection having established a conventional transport protocol;
receiving a plurality of requests from a client of the one or more clients, each request of the plurality of requests having assigned a unique sequence number, each request being an input/output request to a selected data storage device connected to the selected interface to a data storage device;
dividing the plurality of requests into a plurality of subsets of requests;
assigning a batch number to each subset of requests the batch numbers increasing monotonically so that each subset of requests is assigned a unique batch number;
executing a first subset of requests in arbitrary order over the plurality of connections in response to a first batch group number; and
executing a second subset of requests in response to a second batch number after execution of all of the first subset of requests have completed.